[evlatests] More on PDif at 3-bits ...

[Rick Perley]

    A 3-hour test of 3bits were taken last night, using 3C84 (30 Jy), 
two other nearby point sources (2.5 and 0.5 Jy), and blank sky.  The 
goal was to check whether the PDif compression problem affects the 
visibilities.  I predicted it would not.

    Answer:
   
    It definitely does not!   I used the 8-bit data to establish 
accurate flux densities of the two weaker sources (applying TYAPL -- 
there is every indication that the PDifs for 8-bit data are doing what 
they are designed to do -- and no more!).  I then utilized these flux 
densities, and solved for system gains using the 3-bit data.  The 
resulting solutions are an exact mirror image of the PDif values for 
those same data -- in essence, a reflection of the total power seen by 
the synchronous demodulators. 

    I can also report that Barry's conjecture that overflow of the 3-bit 
digitizers is responsible for our PDif troubles is *not* supported by 
comparison of the sampler DC offsets to the PDif compression values.  
Keith extracted these offsets from the monitor data, and I've compared 
them to the PDif depression values -- there is no correlation whatever. 
    Ken worries about sampler rms values -- perhaps a correlation will 
be found there.  I'll do that after lunch.